How BioIVT built agentic research quoting that gives service reps 43% more time for high-impact work

We used Agentforce to build Ivy, a quoting agent that works by reading an opportunity the way a person would — just much faster. 

Ivy looks at what the customer is asking for — like products, quantities, and shipping requirements — and compares that to the more than 15,000 product codes in our Data 360 catalog. It finds the right product, adds it to the opportunity, creates the quote, and routes it to customer service for a quick review before it goes out.

We set it up so that for standard catalog quotes, a case gets assigned to Ivy. There's a minute-long pause before Ivy starts working in case somebody has to interfere or pull it out of the queue for some reason, and then Ivy picks it up and compiles the quote in seconds.

We launched it in July 2025 and the impact has already been pretty dramatic. We’d already accelerated the quoting process when we first moved it into Salesforce, which took us from weeks to days. 

With Agentforce, we went from days to minutes. 

One of our service reps recently shared with me that she used to start every day with around 70 cases in her queue. Daunting, right? Now when she starts her day, Ivy’s already created those quotes for her so all she has to do is review the quotes for accuracy and send them out. 

She has 43% more time for the high-value tasks that really require her expertise, which is exactly what we hoped to accomplish with Ivy.

Honestly, how far we were able to take it. We started with a really narrow use case — creating quotes for simple, standard catalog products — because I didn’t know yet if Ivy could handle the complex stuff. Depending on the research, you might need specific disease states, gender, control samples — there are a lot of variables.

But once we started refining Ivy, we were able to move into more complex, configurable products. That’s where the value really increased, because those are harder for a person to build consistently.

The other surprise was how powerful the feedback loop has been. If a service rep needs to adjust something in Ivy’s quote, we capture that automatically and use it to refine Ivy. Being able to automatically capture changes and use them to improve Ivy over time is how we’ve gotten to our current 94% accuracy rate.

This is where we spent a lot of time. You have to be very clear about what the agent can do — and what it can’t do.

We had to teach Ivy how to interpret the way we work — even down to shorthand like “T2” for type 2 diabetes. And then as we expanded into configurable products, it had to understand things like donor counts and volume per donor.

Setting guardrails was critical. Ivy can’t suggest alternative products so it either finds the right one or escalates to a service rep. In our world, two products that look similar can have very different scientific outcomes and it was imperative that Ivy never be guessing.

Working with Salesforce Premier Success and a prompt engineer made a big difference. They helped us get more structured — how we defined Ivy’s job, how we designed prompts, and what logic should live in flows versus the agent. That clarity is really what improved accuracy and made it something we trust.

Start with a clear, well-defined use case and start small. Let the agent be part of the process, not the whole thing. Put guardrails in place so people can review the output and build trust over time.

Also, be prepared to refine it. You’re going to spend time testing it, adjusting it, and improving it — that’s just part of the process.

And don’t be afraid to push it a little. We use a full sandbox environment so we can test the limits, then come back and decide where we’re comfortable.

If you approach it that way — clear use case, thoughtful rollout, continuous refinement — you’ll get a lot more out of it.

We implemented Salesforce as our core platform to reduce or remove systemwide friction. There was duplicate data entry across systems, we couldn’t really master our account and contact data, and a lot of reporting was happening offline in Excel.

That led to misaligned data and different versions of the truth, which made it hard to operate efficiently. You’d have one team looking at one set of numbers and another team looking at something slightly different.

So a big part of our digital strategy was creating a single source of truth — really getting to a place where we had clean, unified data across the business. We started with sales, then expanded into service, and more recently into marketing, all on the same platform.

Now we can see the full picture of the customer in one place — what they’re buying, what opportunities are in flight, any service issues, and how we’re engaging them.

That unified view is what makes everything else possible — Ivy included. The agent fits right into our digital strategy because it's all on the same platform.

We’ve been very intentional about where data lives and how it flows. Salesforce is our system of record for accounts and contacts, and our ERP is where product codes, pricing, and order fulfillment live.

But the key is that the data between those systems is aligned and connected. We spent a lot of time mastering that — making sure customer IDs, product codes, and transactions all match and move cleanly between systems.

When Ivy creates a quote, there’s a real-time outbound message that goes to the ERP, pulls the correct customer-specific pricing, and sends it right back into Salesforce. It’s happening almost instantly, but it’s only possible because the data is clean and consistent on both sides.

We also built processes so that when a new account or contact is created in Salesforce, it syncs with the ERP and comes back with the right identifiers. That keeps everything tied together from the very beginning.

At the end of the day, it’s that clean, unified data layer — and the integration on top of it — that allows everything to flow from quote to order without breaking.

They were a big part of it. CREO Consulting helped us execute quickly. From our first conversation in March to going live in July, that kind of speed doesn’t happen without the right partners in place.

Building Ivy together really showcased the power of our partnership with Salesforce in action. Salesforce Premier Success gave us access to engineers and experts who helped us work through more complex pieces — like structuring the agent and handling pricing logic. 

And our account team is how we discovered Agentforce in the first place. They’re always bringing us new ideas and were actually the ones who introduced us to the concept of agentic AI as a possible way to help us move faster.

Once we were sure about agentic AI being the right fit, it wasn’t a question of choosing Salesforce over something else. It was more that we already have proven success with Salesforce, so let’s build on it with the partners we trust.